Multiscale deep neural network based analysis of FDG-PET images for the early diagnosis of Alzheimer’s disease

Lu, Donghuan; Popuri, Karteek; Ding, Gavin Weiguang; Balachandar, Rakesh; Beg, Mirza Faisal

doi:10.1016/j.media.2018.02.002

Cited by 166 publications

(143 citation statements)

References 25 publications

Supporting

Mentioning

116

Contrasting

Unclassified

Order By: Relevance

“…The use of two disjoint pipelines and the need to construct ad-hoc kernels can be surmounted by the use of a class of algorithms known as deep learning, which afford much greater representational flexibility than kernelbased methods and also automatically "learn" data transformations which maximize an arbitrary performance metric. Such methods have been applied to AD vs. healthy subject discrimination (Hosseini-Asl et al, 2016;Liu et al, 2018;Payan and Montana, 2015) and pMCI vs sMCI classification (Choi et al, 2018;Lu et al, 2018a, b) As an example, Choi et al, 2018 andLu et al, 2018a use deep learning to achieve one of the highest pMCI/sMCI classification performances to-date ( ~84% -82% conversion rate accuracies for these studies respectively). Their predictions are based on a single (albeit very informative) imaging modality (PET) which employs ionizing radiation.…”

Section: Introductionmentioning

confidence: 99%

A parameter-efficient deep learning approach to predict conversion from mild cognitive impairment to Alzheimer’s disease

Spasov¹,

Passamonti²,

Duggento

et al. 2018

Preprint

View full text Add to dashboard Cite

Some forms of mild cognitive impairment (MCI) can be the clinical precursor of severe dementia like Alzheimer's disease (AD), while other types of MCI tend to remain stable over-time and do not progress to AD pathology. To choose an effective and personalized treatment for AD, we need to identify which MCI patients are at risk of developing AD and which are not. Here, we present a novel deep learning architecture, based on dual learning and an ad hoc layer for 3D separable convolutions, which aims at identifying those people with MCI who have a high likelihood of developing AD. Our deep learning procedures combine structural magnetic resonance imaging (MRI), demographic, neuropsychological, and APOe4 genotyping data as input measures. The most novel characteristics of our machine learning model compared to previous ones are as follows: 1) multi-tasking, in the sense that our deep learning model jointly learns to simultaneously predict both MCI to AD conversion, and AD vs healthy classification which facilitates the relevant feature extraction for prognostication; 2) the neural network classifier employs relatively few parameters compared to other deep learning architectures (we use ~550,000 network parameters, orders of magnitude lower than other network designs) without compromising network complexity and hence significantly limits data-overfitting; 3) both structural MRI images and warp field characteristics, which quantify the amount of volumetric change compared to the common template, were used as separate input streams to extract as much information as possible from the MRI data. All the analyses were performed on a subset of the Alzheimer's Disease Neuroimaging Initiative (ADNI) database, for a total of n=785 participants (192 AD, 409 MCI, and184 healthy controls (HC)). We found that the most predictive combination of inputs included the structural MRI images and the demographic, neuropsychological, and APOe4 data, while the warp field metric added little predictive value. We achieved an area under the ROC curve (AUC) of 0.925 with a 10-fold cross-validated accuracy of 86%, a sensitivity of 87.5% and specificity of 85% in classifying MCI patients who developed AD in three years' time from those individuals showing stable MCI over the same time-period. To the best of our knowledge, this is the highest performance reported on a test set achieved in the literature using similar data. The same network provided an AUC of 1 and 100% accuracy, sensitivity and specificity when classifying NC from AD. We also demonstrated that our classification framework was robust to different co-registration templates and possibly irrelevant features / image sections. Our approach is flexible and can in principle integrate other imaging modalities, such as PET, and a more diverse group of clinical data. The convolutional framework is potentially applicable to any 3D image dataset and gives the flexibility to design a computer-aided diagnosis system targeting the prediction of any medical condition utilizing multi-modal imaging a...

show abstract

Section: Introductionmentioning

confidence: 99%

A parameter-efficient deep learning approach to predict conversion from mild cognitive impairment to Alzheimer’s disease

Spasov¹,

Passamonti²,

Duggento

et al. 2018

Preprint

View full text Add to dashboard Cite

show abstract

“…But in this early stage of this disease the early stage of this dementia the person can be considered as the person who can affected by the dementia disease when take up the symptoms of the disease with clear vision. [20] This early stage of these dementia symptoms may be similar to the symptoms the person affects with 65 years old person.…”

Section: Symptoms and Signsmentioning

confidence: 77%

A Deep Learning Method on Medical Image Dataset Predicting Early Dementia in Patients Alzheimer's Disease using Convolution Neural Network (CNN)

2019

IJRTE

View full text Add to dashboard Cite

Memory loss is one of the major dementia where the human has a common loss of memory which shows the person to behave worst and they don’t care them properly. Alzheimer's disease (Ad) is a neurodegenerative disease which affects the brain with mild cognitive impairment.[4] As MCI has several phases where treatment can be consider for avoiding side effects. Deep Learning techniques is the current trend which can handle the images, massive datasets such as unsupervised, supervised and reinforcement progress.[3] A human MRI images is deal with the existing system to find the dementia. In Existing system 82.51% accuracy of classification of neural network was identified [2][3]. Due to several limitations of existing system CNN was proposed. To predict the dementia an algorithm named Logistic regression is used to produce the accuracy more than a loss function. To the test accuracy betterment OASIS project dataset is utilized.

show abstract

“…Computerized image analysis of PET scans (and also those fused with MRI) for early diagnosis of Alzheimer's disease was reported in the literature [8]- [13]. The main characteristics regarding feature extraction, classification, and image database are summarized and compared in Table I.…”

Section: Introductionmentioning

confidence: 99%

“…The main characteristics regarding feature extraction, classification, and image database are summarized and compared in Table I. In [8], [9], the novel classification methods were focused. Lu et al [8] proposed a novel multiscale deep neural network (MDNN) to learn the patterns of metabolism changes due to AD pathology in FDG-PET images and use them as the discriminant between AD subjects and normal controls (NC).…”

Section: Introductionmentioning

confidence: 99%

“…In [8], [9], the novel classification methods were focused. Lu et al [8] proposed a novel multiscale deep neural network (MDNN) to learn the patterns of metabolism changes due to AD pathology in FDG-PET images and use them as the discriminant between AD subjects and normal controls (NC). Based on the data published by Alzheimer Disease Neuroimaging Initiative (ADNI) the deep architecture performed well, especially in early AD cases.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Classification of Alzheimer’s Disease in PET Scans using MFCC and SVM

Panyavaraporn

Horkaew

2018

Int. J. Adv. Sci. Eng. Inf. Technol.

View full text Add to dashboard Cite

Unlike age-related dementia, Alzheimer's disease is more progressive and causes rapid deterioration in a patient's cognitive functions. Before its first clinical manifestation, it is evident that the damaging brain process has already been commenced much earlier in life. This asymptomatic period could have spanned as long as a decade or more. Although there is not yet an ultimate cure for the disease, the sooner it is diagnosed, the more chance that available therapeutic measures could improve the patient's quality of life. Standard medical questionnaire and medical imaging are the principal means of identifying early Alzheimer's disease. Despite a great effort having been made in analyzing structural atrophy in the human brain by using CT and MRI, the recent attempts have reached high accuracy and precision but relatively poor sensitivity. Functional imaging such as PET is of much lower spatial resolution but promising modality taken to elevate this limitation. This paper presents a classification method for early detection of the disease from PET scans drawn from the Thai population. However, instead of conventional structural analysis, this study performed clustering on unwrapped signals, transformed from imaging data by using Mel-Frequency Cepstral Coefficients (MFCC), by a generic Support Vector Machine (SVM) classifier. The experimental results reported herein indicate that, with the optimal MFCC order, the proposed method could identify subjects with Alzheimer's from controls, with high accuracy, precision, and specificity. With a cross-validation ratio of 8:2 and a linear SVM kernel, the classification accuracy, sensitivity, and specificity were 96.51, 93.98, and 97.77, respectively, and increased as the MFCC orders.

show abstract

Multiscale deep neural network based analysis of FDG-PET images for the early diagnosis of Alzheimer’s disease

Cited by 166 publications

References 25 publications

A parameter-efficient deep learning approach to predict conversion from mild cognitive impairment to Alzheimer’s disease

A parameter-efficient deep learning approach to predict conversion from mild cognitive impairment to Alzheimer’s disease

A Deep Learning Method on Medical Image Dataset Predicting Early Dementia in Patients Alzheimer's Disease using Convolution Neural Network (CNN)

Classification of Alzheimer’s Disease in PET Scans using MFCC and SVM

Contact Info

Product

Resources

About